IEEE TRANSACTIONS ON CONSUMER ELECTRONICS, VOL. 64, NO.

1, FEBRUARY 2018 61

Design and Implementation of an Energy-Saving

Lighting Control System Considering
User Satisfaction
Jinsung Byun, Member, IEEE, and Taehwan Shin, Member, IEEE

Abstract—Lighting consumes the largest amount of energy in characteristics of LED lighting, such as high energy efficiency
buildings. Recently, many studies of energy-efficient lighting sys- and long lifetime.
tems with a variety of sensor and communication technologies In spite of an eco-friendly characteristics of LED lighting,
have been conducted as a way to increase the cost efficiency
of lighting. However, earlier studies have mostly focused on it has still not become widespread, mainly due to its high
energy efficiency, whereas they have not significantly consid- installation costs compared to existing lighting. With recent
ered the occupant’s satisfaction. Therefore, this paper proposes advances in manufacturing technologies, the price of LED
an energy-saving lighting control system considering the occu- lighting is slowly decreasing, whereas the cost of LED lighting
pant’s satisfaction. The proposed system improves the energy is still very expensive. For example, the LED lighting is five
efficiency and the occupant’s satisfaction by controlling lighting
control parameters considering the characteristics of space and times more expensive than the existing lighting devices at cur-
the occupant’s behavior patterns. Moreover, this paper deployed rent market. For this reason, many studies have been widely
the proposed lighting systems in a building and operated them in carried out to enhance the economic feasibility of LED light-
a real work environment to evaluate the performance. The results ing through control algorithms based on sensors and network
showed that the proposed system reduced energy consumption technologies [4]–[15]. These LED lighting systems have been
up to 43% by replacing the existing fluorescent lights with the
proposed lighting control systems. also released in the market. These systems reduce energy by
lowering output power of lighting when the user is not in
Index Terms—Energy efficiency, energy-saving lighting con- the room or when the room is rather too bright, instead of
trol system (ESLiCoS), lighting control system design, user
satisfaction. operating them at the same illuminance.
The most significant problem of existing systems is that
although they can reduce energy, they do not sufficiently
I. I NTRODUCTION consider occupant satisfaction. For this reason, the existing
URRENTLY, in the U.S. energy consumed in buildings systems in the market are used mostly in simple spaces, such
C accounts for about 40% of energy consumption.
Furthermore, about 10% of the energy is consumed by
as restrooms and corridors, but they are not widely used in
complex spaces, such as offices where there are many occu-
lighting for residential buildings and 20% for commercial pants that conduct complicated tasks. Therefore, this paper
buildings. Lighting is one of the most energy consumed seg- proposes an energy-saving lighting control system (ESLiCoS)
ment in a building. Thus, it is most important to reduce considering user satisfaction. The ESLiCoS improves energy
the energy consumed by lighting in buildings. Ongoing efficiency and occupant satisfaction by controlling lighting
efforts have been made to reduce the energy consumed control parameters considering the characteristics of a space
by lighting. Especially, the development of a light-emitting and the occupant behavior patterns. The ESLiCoS is developed
diode (LED) has affected significant changes in the lighting and evaluated, focusing on buildings. The proposed systems
industry [1]–[3]. Governments of various countries are provid- are deployed in the test bed and operate in a real work
ing various support programs for the spread of LED lighting. environment to analyze and evaluate energy efficiency.
For example, the South Korean government is making efforts The remainder of this paper is organized as follows.
for the spread of LED lighting by increasing the LED lighting Sections II and III describe the related works and research
supply budget, the mandatory installation of LED lighting in methodology. Section IV presents the proposed lighting con-
public buildings, and providing subsidies. The gradual changes trol system (ESLiCoS) and addresses design techniques and
in the lighting industry have been caused by the eco-friendly guidelines for the lighting control system to enhance energy
efficiency and user satisfaction. Section V presents the deploy-
Manuscript received January 1, 2018; revised February 10, 2018; accepted ment and experiment on energy consumption. Finally, the
February 15, 2018. Date of publication March 7, 2018; date of current version conclusion is given in Section VI.
March 29, 2018. (Corresponding author: Taehwan Shin.)
J. Byun is with the Analysis and Evaluation Division, Defense
Agency for Technology and Quality, Jinju 52851, South Korea (e-mail:
[email protected]). II. R ELATED W ORKS
T. Shin is with the National IT Industry Promotion Agency, Jincheon-gun
27872, South Korea (e-mail: [email protected]). Many studies of lighting control for energy savings have
Digital Object Identifier 10.1109/TCE.2018.2812061 been carried out. First, there are many papers about lighting
1558-4127 c 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
62 IEEE TRANSACTIONS ON CONSUMER ELECTRONICS, VOL. 64, NO. 1, FEBRUARY 2018

control according to the user’s preference and the user’s profile

to reduce energy consumption. Pan et al. [8] proposed a light
control system based on wireless sensor networks (WSNs).
An illuminance decision algorithm and a device control algo-
rithm are devised to meet the users’ illuminance requirements
according to their activities and profiles. Wang and Tan [16]
developed a solution for illuminance control of a LED light-
ing system by using a neural network model based on the Fig. 1. Basic operating principal of the proposed system.
mapping relationship between the light dimming level and the
table illuminance. The objective of this paper is to minimize
energy consumption of the overall lighting system and provide control by sensing user’s movement and daylight condi-
personalized lighting control services by optimizing the illumi- tions. However, these studies and products mostly focused on
nance levels. Bhardwaj et al. [14] proposed indoor solid state energy efficiency; they did not significantly consider occupant
lighting using LED based on illumination model. The pro- satisfaction.
posed system dynamically adjusted light brightness according
to the user’s preferences. Wen and Agogino [17] presented III. S TUDY M ETHODOLOGY
a dynamic personalized optimal lighting design method capa- The proposed system basically utilizes a motion sensor and
ble of adjusting the illuminance of a light according to each an illuminance sensor to recognize surrounding environments.
occupant’s preference. Yun et al. [18] conducted a field sur- It controls the illuminance of LED lighting based on energy-
vey on occupancy, lighting use patterns, and lighting energy saving mechanisms proposed in this paper. When designing
consumption. This paper showed that the lighting use patterns an LED lighting system, the existing products and studies
are related to the occupancy patterns, and users have a ten- have mainly focused on the energy saving aspect, whereas
dency to switch on the lighting when they first arrive. Second, this paper focuses on the user satisfaction aspect as well as
many studies of lighting control considering daylight have the energy saving aspect. Since an autonomous change in the
been conducted. Matta and Mahmud [6] presented a light con- illuminance of lighting causes the significant inconvenience
trol system that uses a controller area network (CAN). A logic to users, the user satisfaction is a very important issue that
method is introduced to control blind angles in order to uti- must be addressed. Hence, this paper designs and evaluates
lize daylight. Bai and Ku [7] proposed automatic room light an intelligent lighting control system considering both the
intensity detection and light control scheme. In addition, many energy-saving aspect and the user satisfaction aspect.
studies have focused on reducing the power consumption of On the other hand, building consist of various spaces, such
lighting by taking into account the daylight conditions and the as offices, conference rooms, hallways, stairs, lounges, parking
user’s patterns simultaneously. Pandharipande and Caicedo [9] lots, etc. Since each space has different characteristics, differ-
introduced the energy-efficient illumination control design of ent lighting control policies need to be applied according to the
an LED-based lighting system that considers the location of characteristics. This method can improve the performance of
occupants and daylight distribution. An ultrasound-based array energy saving and user satisfaction simultaneously. Therefore,
sensor was utilized in order to determine the location of occu- this paper addresses the method for controlling LED lighting
pants. Park and Hong [10] proposed a reference model for according to the characteristics of spaces. Spaces in buildings
a lighting control system based on a building automation and are classified into three broad categories: 1) wide type spaces;
control network with consideration of the daylight conditions. 2) small type spaces; and 3) corridor type spaces.
Finally, there are many papers about lighting control based on This paper implemented the proposed system. Six light-
the user’s movement. Uhm et al. [11] presented a power-aware ing control parameters were used for lighting control. The
LED light enabler that is equipped with an illuminance sensor, proposed lighting control systems are deployed in the test
a motion sensor, and a network interface. The presented system bed and operated them in a real work environment. Energy
reduces energy consumption by controlling lighting depending efficiency and user satisfaction were evaluated by analyzing
on user’s movement and the brightness of a room. In addition, data collected from the deployed systems and occupants in
Bellido-Outeirino et al. [5] presented a remote management the test bed.
system for building lighting automation with integration of
digital addressable lighting interface (DALI) devices in WSNs.
IV. E NERGY-S AVING L IGHTING C ONTROL S YSTEM
The main goal of DALI is to provide users with a fully cen-
C ONSIDERING U SER S ATISFACTION
tralized building automation service that combines heating,
ventilating, air-conditioning, and lighting service. DALI can A. ESLiCoS Control Parameters
reduce installation and maintenance costs through centralized The proposed system, ESLiCoS, controls the brightness
management. of lighting based on the six lighting control parameters.
Currently, a lighting control system that considers user’s Fig. 1 illustrates the basic operating principles of the pro-
preferences, daylight conditions, and user’s movement patterns posed system based on these lighting control parameters. The
for energy savings is being extensively studied in academia and definitions of the control lighting parameters are as follows.
industry. Furthermore, some companies are introducing intelli- 1) Lmax : Maximum light brightness (illuminance, lux).
gent lighting devices to the market that are capable of dimming 2) Lmin : Minimum light brightness (illuminance, lux).
BYUN AND SHIN: DESIGN AND IMPLEMENTATION OF ESLiCoS CONSIDERING USER SATISFACTION 63

3) Tr : Time taken for light brightness to increase.

4) Tm1 : Time between when the movement is not detected
(when the light intensity is Lmax ) and when the light
brightness begins to reduce.
5) Tm2 : Time between when the movement is not detected
(when the light intensity is Lmin ) and when the light
brightness begins to reduce.
6) Tf : Time taken for the light brightness to reduce.
This paper uses the technique of properly adjusting the
light control parameters. The proposed system maintains Lmin
though user’s movement is not detected, such as during Tm2 .
Fig. 2. FSM description of autonomous light control in wide type spaces.
Lmin can of course be set. In spaces where user movement is
not frequently detected, such as parking lots and warehouses,
Lmin is set to low; while in spaces where user movement is fre-
a user reads a book, and the light is darkened and brightened
quently detected, such as offices and lobbies, Lmin is set to high
repeatedly, we can expect that the user will feel annoyed and
in order to minimize user’s inconvenience due to rapid change
cannot continue to read a book in a comfortable state. These
in the illuminance of lights. If the proposed system detects
examples imply the following two points.
user’s movement, the illumination of lighting reaches Lmax for
1) The number of times the illuminance becomes Lmin
Tr . The length of Tr can also be set, and if the length is too
(when the illuminance is Lmax ) or the illuminance
short, users feel uncomfortable due to the rapid change in illu-
becomes Lmax (when the illuminance is Lmin ) is closely
mination. Empirically, it is acceptable to set Tr to about 1.0 s
related to the occupant’s satisfaction. In other words, the
−3.0 s. If the illuminance reaches Lmax , three types of rules
more number of times the illunimance becomes Lmin or
will be followed. First, if no movement is detected, the system
Lmax , the more uncomfortable the user feels.
starts the countdown timer (Timer = Tm1 ). Second, if a move-
2) The difference in the illuminance between Lmax and
ment is detected, the countdown timer is reset. Third, if the
Lmin is closely related to the occupant’s satisfaction. In
countdown timer reaches a timeout (Timer = 0), Lmax is slowly
other words, the larger the difference in the illuminance
reduced to Lmin . The slow reduction in the illuminance is to
between Lmax and Lmin is, the more uncomfortable the
give visual comfort to users. When the illuminance reaches the
user feels.
Lmin , the countdown timer is initially started (Timer = Tm2 ).
Therefore, this paper proposes a new lighting control system
At this time, if a user movement is detected, the countdown
considering both energy efficiency and user satisfaction
timer is stopped and the illuminance is increased up to Lmax
for Tr ; the countdown timer is restarted. When the illuminance
is Lmin , if no user movement is detected continuously and the C. Adaptive Light Control Schemes Based on the
countdown timer reaches timeout, Lmin is slowly reduced to Characteristics of the Space
zero. This paper will discuss various issues about the deter- When a light is autonomously controlled according to the
mination of the lighting control parameters. Lmax setting is user’s movement, the most significant problem is that users
also important in order to save energy. In many cases, the are likely to feel very uncomfortable due to the frequent
existing lighting systems installed in a building are too bright. changes in the brightness of the light. For example, a user
This paper will also discuss the need of Lmax control through walks along a quiet corridor, where the autonomously con-
lighting simulation to reduce energy consumption. trolled lighting systems are installed. The user would be
walking in a dark corridor because the light turns on only
if the user is close to the light. To solve this problem,
B. Problem Description and Design Goals the proposed system turns the light down and maintains
The lighting control parameters (i.e., Lmin , Lmax , Tr , Tm1 , the constant brightness (Lmin ) without turning the light off
Tm2, and Tf ) are closely related to the amount of energy sav- completely.
ings. If user movement is detected, the illuminance increases Lmin has a significant impact on both energy savings and
up to Lmax , whereas if user movement is not detected, the occupant satisfaction. For example, if Lmin is high, users will
illuminance decreases up to Lmin . This implies that user’s not feel uncomfortable, but the amount of energy reduction
inconvenience could increase because of the frequent change will be relatively small. Similarly, if Lmin is low, the amount
of light’s illuminance. Similarly, as Lmax and Lmin are reduced, of energy reduction will be significantly large, but users are
the amount of energy reduction becomes large, whereas this prone to feel uncomfortable.
implies that user’s inconvenience could increase because of the In addition, it can cause the significant waste of energy to
dark indoor environment. Therefore, it is necessary to prop- control lights installed in different spaces based on same pol-
erly determine the lighting control parameters in consideration icy. For example, energy is wasted when a light maintains high
of the relationship between the energy saving effect and the Lmin in spaces where the user’s movement is rarely detected.
user’s inconvenience. Users tend to feel the greatest inconve- Therefore, this paper presents an adaptive light control scheme
nience when they readily recognize changes in the illumination based on the characteristic of the space in a building. This
while lighting is autonomously controlled. For example, when paper takes advantage of the different properties of spaces in
64 IEEE TRANSACTIONS ON CONSUMER ELECTRONICS, VOL. 64, NO. 1, FEBRUARY 2018

Fig. 3. FSM description of autonomous light control in corridor type spaces.

a building in order to improve energy efficiency and occupant

satisfaction.
A building consists of various spaces, such as parking lots,
lobbies, elevators, stairs, corridors, private offices, seminar
rooms, restrooms, break rooms, and warehouses. It is possible
for these spaces to be classified into three groups: 1) wide
type spaces; 2) small type spaces; and 3) corridor type spaces.
These classifications can be used for controlling lights effec-
tively. In other words, this paper controls lights by applying Fig. 4. FSM description of autonomous light control in small type spaces.
different lighting control policies to the classified spaces in
order to enhance energy efficiency and occupant satisfaction. 3) Autonomous Light Control in Small Type Spaces: In
Basically, this paper utilizes a group-based control method small type spaces, since the number of users managed by the
of which the proposed system simultaneously controls the system is small, it is possible to exactly determine the pres-
grouped lights that are grouped in a circle around a user in ence or absence of users without error. Therefore, in small
order to reduce user’s discomfort more. type spaces, if the system confirms that there is no user, the
1) Autonomous Light Control in Wide Type Spaces: system turns the lights off, which does not make the users
Fig. 2 illustrates the finite state machine (FSM) description of uncomfortable. In other words, in case of small type spaces,
autonomous light control in wide type spaces. In the wide type if there is no user, the proposed system does not maintain
spaces, the proposed system uses a two-step control process. the constant illuminance, Lmin , but turns the light completely
If user movement is detected, the system increases the illumi- off. When a user enters the space, the system increases the
nance up to Lmax and sends a control packet (L = Lmax ) to the light illuminance up to Lmax . If it is confirmed that the user
systems belonging to the same group in order to make them leaves the space, the system decreases the illuminance to zero.
increase the illuminance up to Lmax . The system receiving this Fig. 4 shows the FSM description of autonomous light control
control packet also increases the brightness up to Lmax . After in small type spaces. In additon, in the FSM of autonomous
increasing the brightness up to Lmax , the system starts a count- light control in wide type spaces, if Lmin is set to zero, the
down timer that can interrupt the system after a given amount same lighting control effect can be obtained.
of time has expired.
2) Autonomous Light Control in Corridor Type Spaces:
In corridor type spaces, user’s movement patterns are pre- D. Guidelines for the Determination of Lmin
dictable. By taking advantage of this typical feature, energy Table I shows guidelines of the Lmin setting based on the
can be saved further while reducing the user’s feeling of dis- spaces. Lmin can be set depending on the complexity of activi-
comfort. Darkening the lights located in the direction of user ties and the frequency of space utilization. For example, when
movement causes considerable user’s discomfort. Therefore, the frequency of space utilization and the complexity of activ-
the ESLiCoS controls the lights by distinguishing between ities are high, setting Lmin high reduce discomfort. This paper
the lights in the opposite direction of the user’s movement defines the complexity of activities as the concentration of
(Ldirect = back) and the lights located in the direction a user, which refers to a General Social Survey on Time Use
of user movement (Ldirect = front) in order to improve by Statistics Canada that determines the complexity of the
energy efficiency and user satisfaction. Fig. 3 shows the representative activities as described below.
FSM description of autonomous light control in corridor type 1) Activity of High Complexity: Analyzing, reading, writ-
spaces. ing, studying, activities related to paid work.
BYUN AND SHIN: DESIGN AND IMPLEMENTATION OF ESLiCoS CONSIDERING USER SATISFACTION 65

TABLE I
G UIDELINES FOR THE D ETERMINATION OF Lmin Fig. 6 shows the simulation results. The expected bright-
ness of a space can be used as guidelines for determining
Lmax before replacing a fluorescent light with an LED light.
The guidelines for office lighting design issued by Seoul
Metropolitan Government recommend that about 400lx-500lx
is appropriate for typing, reading, writing, and meeting. Hence,
it is recommended that the existing 55W LED light in the test
is replaced by a 35W or 42W LED light to save energy.

F. Implementation of the ESLiCoS

In this paper, the prototype of the ESLiCoS is implemented
by applying the proposed schemes and algorithms.
1) Hardware Implementation: Fig. 7 shows the prototype
of the proposed system. The ESLiCoS uses an 8-bit micro-
controller as the main processor part. This part plays a role in
the situation analysis, event processing, and learning. Also, it
optimizes the control parameters to adapt to the various envi-
ronments. To provide energy saving services, two kinds of
sensors (i.e., a movement detection sensor and an illuminance
sensor) are basically required. A ZigBee (250 kbps/2.4 GHz)
module is used for communication with other LED lighting
systems and networked devices. LED driver part consists of
current controller modules for driving LEDs. The power part
is composed of a power regulator and a switched-mode power
supply.
Fig. 5. Illuminous intensity distribution pattern of the light for simulation. 2) Middleware Implementation: The existing LED light-
ing control systems are mainly developed using a low-cost
TABLE II
C ONFIGURABLE PARAMETERS AND T HEIR P REDETERMINED VALUES
microcontroller unit in order to reduce unit production costs.
Thus, they have limited capability of computing and stor-
age resources. To solve this problem, this paper designed an
adaptive middleware platform that can optimize middleware
architecture according to the changes in the external envi-
ronments. The adaptive middleware platform is composed of
the LED control module group (which performs the role of
controlling the LED), the adaptive middleware group (which
is optimized according to the changes in the external envi-
ronments), and the table group (which manages a variety of
data used for context awareness and LED control).

V. D EPLOYMENT AND E XPERIMENT

A. Test Bed for Performance Evaluation
This paper evaluated the performance of the proposed light-
2) Activity of Middle Complexity: Cooking, washing, activ- ing control system by using the test bed. The ESLiCoSes are
ities related to unpaid work. installed and operated in a real work environment. In this
3) Activity of Low Complexity: Walking, parking, sleeping, paper, adjusted Lmax is defined as Lmax with reduced bright-
eating, resting, watching TV, toileting. ness via pulse width modulation control. For example, if Lmax
is set to adjusted Lmax , then when the ESLiCoS detects the
user’s movement, the light brightness will be high up to this
E. Guidelines for the Determination of Lmax value. The adjusted Lmax was set differentially with reference
This paper simulated the brightness of the test bed when to the result of the light brightness simulation.
the existing fluorescent light was replaced with a 35W, a 42W, Table III shows the lighting control systems installed in the
a 50W, and a 55W LED light by using the light simulation test bed. In the table, simulation brightness refers to the aver-
tool as shown in Fig. 6. The simulation was carried out using age brightness of the space that is calculated by the lighting
the lights that is installed in the actual building. The config- simulation. According to the simulation brightness, this paper
urable parameters and their predetermined values are shown sets the ratio of adjusted Lmax to maximum Lmax in order
in Table II. Fig. 5 shows the illuminous intensity distribution to make the brightness of the space suitable with reference
pattern of the LED light used for simulation. to the guidelines for office lighting design issued by Seoul
66 IEEE TRANSACTIONS ON CONSUMER ELECTRONICS, VOL. 64, NO. 1, FEBRUARY 2018

Fig. 6. Results of the light brightness simulation.

Fig. 8. Comparison of energy consumption.

Fig. 7. Prototype of the proposed system.

Metropolitan Government. These guidelines show that approx-

imately 400lx-500lx is appropriate to the office for typing,
reading, writing, and meeting, and about 300 lx-500 lx to the
corridor and stairs. This paper also sets the value of Lmin to
50% of maximum Lmax .

B. Results of Experiments on Energy Savings

To evaluate the energy saving performance of the ESLiCoS,
this paper measures the energy consumption in the offices,
corridors, and stairs where the existing 58 W fluorescent
Fig. 9. Analysis of the result of experiment on energy saving.
lights are installed for 20 days. And then, the existing flu-
orescent lights are replaced with the ESLiCoS (LED, 42W),
and measures the energy consumption under the same con-
ditions for 20 days. Fig. 8 shows the experimental results. light (27.6%), a reduction by adjusting Lmax depending on the
The proposed system reduced energy consumption of 43% by results of the light brightness simulation (6%), and a reduction
replacing the existing fluorescent lights. The energy savings by autonomously controlling lights according to the occupant’s
include a reduction by replacing the 58W light with the 42W movement (9.4%) as shown in Fig. 9.
BYUN AND SHIN: DESIGN AND IMPLEMENTATION OF ESLiCoS CONSIDERING USER SATISFACTION 67

TABLE III
I NTELLIGENT L IGHTING C ONTROL S YSTEMS VI. C ONCLUSION
I NSTALLED IN THE T EST B ED This paper designed and developed an intelligent lighting
control system, called ESLiCoS considering energy efficiency
and occupant satisfaction. The ESLiCoS is equipped with
a movement detecting sensor and reduces energy consumption
by autonomously controlling a light based on the occupant’s
movement. The ESLiCoS improved energy efficiency and
occupant satisfaction by controlling lights considering the
characteristics of a space and the occupant behavior pat-
terns. This paper also presented guidelines on the design
of an energy-saving lighting control system considering user
satisfaction. Furthermore, the ESLiCoSes were installed in
a building and operated in a real work environment to evalu-
ate the performance. This paper measured energy consumption
in offices, corridors, and stairs, where the 58W fluorescent
lights were installed for 20 days. Then, the existing fluo-
rescent lights are replaced with the ESLiCoS (LED, 42W),
and measured energy consumption under the same experimen-
tal conditions for 20 days. The test results showed that the
ESLiCoS reduced energy consumption by 43%. The energy
saving results include a reduction by replacing the 58W flu-
orescent light with the 42W LED light (27.6%), a reduction
by adjusting Lmax depending on the results of the light bright-
ness simulation (6%), and a reduction by controlling the light
intelligently according to the movement of occupants (9.4%).
As a further work, the project is planned to develop lighting
control policies according to time variation and the types of
work in a space, and to develop energy saving models based
on the occupant behavior patterns and the values of the light
control parameter.

R EFERENCES
[1] J.-C. Hsieh, D. T. W. Lin, and C.-H. Cheng, “Optimization of ther-
mal management by integration of an SCGM, a finite-element method,
and an experiment on a high-power LED array,” IEEE Trans. Electron
Devices, vol. 58, no. 4, pp. 1141–1148, Apr. 2011.
[2] N. L. Ploch et al., “Effective thermal management in ultraviolet light-
emitting diodes with micro-LED arrays,” IEEE Trans. Electron Devices,
vol. 60, no. 2, pp. 782–786, Feb. 2013.
[3] H. Yang, T. C. W. Schenk, J. W. M. Bergmans, and A. Pandharipande,
“Enhanced illumination sensing using multiple harmonics for LED
lighting systems,” IEEE Trans. Signal Process., vol. 58, no. 11,
pp. 5508–5522, Nov. 2010.
[4] J. Byun, I. Hong, B. Lee, and S. Park, “Intelligent household
LED lighting system considering energy efficiency and user satis-
faction,” IEEE Trans. Consum. Electron., vol. 59, no. 1, pp. 70–76,
Feb. 2013.
[5] F. J. Bellido-Outeirino, J. M. Flores-Arias, F. Domingo-Perez,
A. Gil-de-Castro, and A. Moreno-Munoz, “Building lighting automa-
tion through the integration of DALI with wireless sensor net-
works,” IEEE Trans. Consum. Electron., vol. 58, no. 1, pp. 47–52,
Feb. 2012.
The difference between the ESLiCoS and existing lighting [6] S. Matta and S. M. Mahmud, “An intelligent light control system for
control system is that the ESLiCoS takes into consideration power saving,” in Proc. Annu. Conf. IEEE Ind. Electron. Soc., Glendale,
AZ, USA, Nov. 2010, pp. 3316–3321.
the user’s satisfaction. In other words, the existing studies and [7] Y.-W. Bai and Y.-T. Ku, “Automatic room light intensity detection and
products mostly focused on energy efficiency. That is, they control using a microprocessor and light sensors,” IEEE Trans. Consum.
did not significantly consider the occupant’s satisfaction. This Electron., vol. 54, no. 3, pp. 1173–1176, Aug. 2008.
[8] M.-S. Pan, L.-W. Yeh, Y.-A. Chen, Y.-H. Lin, and Y.-C. Tseng,
paper also conducted a questionnaire survey of 259 occupants “A WSN-based intelligent light control system considering user activ-
who actually work in the test bed. The results showed that 79% ities and profiles,” IEEE Sensors J., vol. 8, no. 10, pp. 1710–1721,
of those surveyed did not feel uncomfortable. Furthermore, Oct. 2008.
[9] A. Pandharipande and D. Caicedo, “Daylight integrated illumination
28% of those surveyed did not even recognize the fact that control of LED systems based on enhanced presence sensing,” Energy
the lighting was controlled. Build., vol. 43, no. 4, pp. 944–950, Apr. 2011.
68 IEEE TRANSACTIONS ON CONSUMER ELECTRONICS, VOL. 64, NO. 1, FEBRUARY 2018

[10] T.-J. Park and S.-H. Hong, “Experimental case study of a BACnet-based Jinsung Byun (M’14) received the B.S. and Ph.D.
lighting control system,” IEEE Trans. Autom. Sci. Eng., vol. 6, no. 2, degrees from the School of Electrical and Electronic
pp. 322–333, Apr. 2009. Engineering, Chung-Ang University, Seoul, South
[11] Y. Uhm, I. Hong, G. Kim, B. Lee, and S. Park, “Design Korea, in 2008 and 2014, respectively.
and implementation of power-aware LED light enabler with He was a Research Professor with Information
location-aware adaptive middleware and context-aware user pat- Technology Research Center, Industrial Security
tern,” IEEE Trans. Consum. Electron., vol. 56, no. 1, pp. 231–239, Research Center, Chung-Ang University supported
Feb. 2010. by the Ministry of Science, ICT and Future
[12] J. Zhang and H.-J. Kim, “Design of smart LED lighting switch Planning in 2014. From 2014 to 2016, he was an
with learning user’s light controlling pattern,” in Proc. Int. Conf. ICT Assistant Professor with the Division of Computer
Converg., Seoul, South Korea, Sep. 2011, pp. 656–657. Engineering, Dongseo University, Busan, South
[13] Y. K. Tan, T. P. Huynh, and Z. Wang, “Smart personal sensor Korea. He is currently a Senior Researcher with the Defense Agency for
network control for energy saving in DC grid powered LED light- Technology and Quality, Jinju, South Korea. His current research inter-
ing system,” IEEE Trans. Smart Grid, vol. 4, no. 2, pp. 669–676, ests include Internet of Things, context-aware service platform, data min-
Jun. 2013. ing, home/building energy management system, ubiquitous computing, and
[14] S. Bhardwaj, T. Ozcelebi, R. Verhoeven, and J. Lukkien, “Smart indoor situation-aware technologies.
solid state lighting based on a novel illumination model and implemen-
tation,” IEEE Trans. Consum. Electron., vol. 57, no. 4, pp. 1612–1621, Taehwan Shin (M’11) received the M.S. and Ph.D.
Nov. 2011. degrees from the School of Electrical and Electronics
[15] Y. S. Cho, J. Kwon, S. Choi, and D.-H. Park, “Development of smart Engineering, Chung-Ang University, Seoul, South
LED lighting system using multi-sensor module and Bluetooth low Korea, in 2012 and 2018, respectively.
energy technology,” in Proc. IEEE Int. Conf. Sens. Commun. Netw., From 2012 to 2015, he was a Senior Research
Singapore, Jun. 2014, pp. 191–193. Staff with the National IT Industry Promotion
[16] Z. Wang and Y. K. Tan, “Illumination control of LED systems based Agency (NIPA), Cloud Computing Support Center,
on neural network model and energy optimization algorithm,” Energy Seoul. From 2015 to 2018, he was a Research
Build., vol. 62, pp. 514–521, Jul. 2013. Staff with the Intelligent Systems and Human Robot
[17] Y.-J. Wen and A. M. Agogino, “Personalized dynamic design of net- Laboratory, Chung-Ang University, Seoul. He is cur-
worked lighting for energy-efficiency in open-plan offices,” Energy rently a Senior Research Staff with NIPA, Jincheon-
Build., vol. 43, no. 8, pp. 1919–1924, Aug. 2011. gun, South Korea. His current research interests include the design of
[18] G. Y. Yun, H. Kim, and J. T. Kim, “Effects of occupancy and lighting embedded systems, intelligent system, architecture design, nonlinear con-
use patterns on lighting energy consumption,” Energy Build., vol. 46, trol, neural networks, Internet of Things, and middleware architecture of
pp. 152–158, Mar. 2012. context-aware system.